Encounter Form for Patients with Acute Knee Injuries

TO THE EDITOR: Thank you for the “Point-of-Care Guide”¹ in the March 15, 2005, issue of American Family Physician. I have several comments that I hope are helpful:

Although the pivot shift maneuver is probably the most accurate physical examination for detecting a torn anterior cruciate ligament (ACL) in the chronic setting,² it is not likely to be very useful acutely. Patients with an acutely torn ACL usually have a significant hemarthrosis with associated guarding and decreased range of motion. These factors make the pivot shift maneuver difficult to perform (and uncomfortable for the patient) in the acute setting.³ Clearly, the Lachman maneuver is a more appropriate test for ACL integrity acutely.

The likelihood ratios and predictive values used in the encounter form¹ are derived from only one study.⁴ They seem at odds with my clinical experience and also with the results of a prior meta-analysis.² The study⁴ by Jackson and colleagues discounted patient history as a way of determining pretest probability, and instead relied on estimated prevalence of the particular injuries in the population. Although the evidence-based literature has not proven the utility of the history in diagnosing acute knee injuries, I believe history is the key to accurate diagnosis. For example, in the case mentioned in the article¹ regarding the author’s knee injury, the most obvious diagnosis, based on history alone, was clearly a torn (most likely, a bucket-handle tear) meniscus with intermittent locking caused by displacement of the flap. The physical examination did not appear to contribute much to the diagnosis, although the presence of a small effusion would be consistent with a meniscal tear.

Joint line tenderness has been found to be the most sensitive indicator of a meniscal tear, although it has poor specificity.^2,5,6 McMurray’s maneuver has been found to have a relatively low sensitivity, with varying specificity,^2,6 and also is extremely difficult to perform in the setting of an acute knee injury. If the patient is in pain, has a significant effusion, has decreased range of motion, or is guarding, the maneuver cannot be performed accurately. Further, it will be painful for the patient, and may prevent the patient from relaxing for any other part of the examination. Various modifications (e.g., Apley’s) may be used in this setting but also have been found to have relatively poor accuracy.^2,6 Newer maneuvers, such as the Thessaly test, may be useful and merit further evaluation.

The patient encounter form¹ is very concise. However, a primary care physician who uses this form and does not have a good musculoskeletal medicine background may miss or not even consider other important knee injuries, including damage to the posterior cruciate ligament, collateral ligaments, posterolateral corner, and patellofemoral joint. Also, the form does not prompt the user to evaluate or record range-of-motion, strength, or neurovascular deficits. No form is perfect, but perhaps adding a few items in this case might be worthwhile.

IN REPLY: I would like to thank Dr. Cohen for his thoughtful comments regarding the encounter form for acute knee injury.¹ I agree with his comments about the pivot shift and McMurray tests, which may be difficult to perform in patients who have a great deal of pain and swelling. Although the form does state “Recheck in ___ days,” it would be more clear to add an item reading “Examination limited; reexamine in ___ days,” to the Assessment/Plan portion of the encounter form. I also have added prompts reminding users to evaluate the patient’s strength, range of motion, and neurovascular status. This revised version of the encounter form is available online at https://www.aafp.org/afp/2005/0315/p1169.html.

The data regarding the accuracy of individual elements of the knee examination were based largely on the meta-analysis by Jackson and colleagues published in 2003.² This study agrees with Dr. Cohen’s comment that joint line tenderness is sensitive but not specific, and that the opposite holds true for the McMurray test. Dr. Cohen cites a study³ performed by a single physician in Turkey on Turkish military recruits 18 to 20 years of age. The author³ found a very good sensitivity and specificity of this finding for lateral meniscal tears, which was better than had been reported previously in the literature. Because this study³ is only one unblinded study based on the physical examination of one examiner, it is of a lower level of evidence than the well-done meta-analysis by Jackson.²

Regarding the differences between the results of Jackson’s 2003 meta-analysis² and Solomon’s 2001 meta-analysis,⁴ in some cases they are small and clinically unimportant. For example, the positive and negative likelihood ratios for joint line tenderness were 1.1 and 0.8, respectively, in Jackson; and 0.9 and 1.1 in Solomon. When they differ, as with the McMurray test, which Jackson found to be more accurate than did Solomon, it may be because Jackson’s article identified an additional study. Ultimately, I chose to rely on data from the Jackson meta-analysis because it was more recent and used what I thought were the more rigorous methods.

Finally, I did not intend to minimize the role of the history, and did include several key findings as “checkoffs.” Unfortunately, existing studies^5,6 have not shown that the clinical history is highly accurate for distinguishing between different kinds of knee injury. However, these studies^5,6 are limited in number and quality. I think there is a great deal more work to be done in developing more rational strategies for evaluating patients with acute knee injury.